[AMBARI-11752] Kerberos: adjust ambari headless principals for unique names

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.1.0
• Fix Version/s: 2.1.0
• Component/s: ambari-server
• Labels: kerberos kerberos_descriptor
• Assignee: Robert Levas
• Reporter: Robert Levas
• Created: 06/Jun/15 10:31
• Updated: 08/Jun/15 17:02
• Resolved: 08/Jun/15 14:33

Description

1) Rollup all headless principal names up to Ambari Principals tab. Currently looks like Storm and Spark are on second tab, under their section, not under Ambari tab with ambari-qa, hdfs, hbase, etc. Also make sure the UI has user readable labels like the others for consistency (see the screen shot. spark.history.kerberos.principal should be "Spark user principal" for example).

2) By default, all of these to be cluster-name scoped by default.

-${cluster_name}

It does no harm for those that don't care... And for those that care about headless principal names to be unique, this ends up being done by default (and saves the user from having to remember to set it this way).

Ultimately when users want to add variables to their principal names they will be doing it across the board - whatever we can do to make it easier for users to do so, would be better. If we had all principals in one pane they can quickly add all of them and visually validate.

Solution
Update the details for all user (identities/type = user) Kerberos Identity entries in kerberos.json files to add the following to the principal name

-${cluster_name}

For example:

${hadoop-env/hdfs_user}@${realm}

to

${hadoop-env/hdfs_user}-${cluster_name}@${realm}

Issue Links

• links to: ReviewBoard